(2010): Debris-flow activity along a torrent in - Dendrolab.ch

Author's personal copy
A. Sorg et al. / Dendrochronologia 28 (2010) 215–223 221
Fig. 5. Germination dates of the sampled trees suggest that a devastating
debris-flow event in the 1880s had cleared the surface and
scoured the currently active channel. This interpretation is supported
by topographical maps from 1881 to 1891, which show a dislocation
of the channel (Swisstopo, 2006).
time elapsed between tree germination and the time needed
for the trees to reach sampling height, we added a locally
observed average tree age at sampling height, thus respecting
accelerated juvenile growth.
The time elapsing between the deposition of material by a
debris flow and the actual colonization of the bare surfaces by
tree seedlings varies between a single year and several years
up to decades (Yoshida et al., 1997; Koch, 2009). This uncertainty
cannot be reduced further, as it is an inherent element
of the methodology. At our study site, there is clear observational
evidence that recently cleared surfaces are recolonized
by pioneer species such as L. decidua and Betula pendula
Roth. within only a few years after geomorphic events (personal
observations and L. Jörger, verbatim). This rather short
germination lag time (ecesis) is most likely due to favourable
climatic and soil conditions such as low elevation and substantial
input of fine-textured material, but it is also the result
of abundant seed sources in nearby forests.
Based on the germination dates of the oldest trees sampled
along the channel, we suggest that a devastating debris flow
must have occurred in the early 1880s, eliminating the former
stand in the area of the current channel. This interpretation
is supported by two topographical maps dating back to 1881
and 1891 (Fig. 5, Swisstopo, 2006), where a dislocation of
the debris-flow channel on the cone is evident, changing its
course from an east–west to the current southeast–northwest
direction. Such drastic changes in the flow direction are often
associated with major debris-flow activity, which apparently
occurred in the Geisstriftbach torrent sometimes between
1881 and 1891.
In addition, debris flows strongly determine individual
tree growth and stand dynamics. For trees along the Geisstriftbach
torrent, we propose that succession is hampered
repeatedly by debris flows in the upper part of the channel,
and that trees are fairly easily removed by incidents. In contrast,
debris-flow activity appears to have a less devastating
and disturbing effect on trees with increasing distance from
the cone apex, where slope gradients are considerably smaller
and the impacts of individual debris-flow surges less severe.
Besides the fact that trees are much younger in the upper part
with germination dates between 1930 and 1970, they have
also been disturbed more frequently. Close to the cone apex,
only a few L. decidua were able to colonize and broadleaved
pioneer trees such as Sorbus aucuparuia L. and B. pendula
are dominant today. Thus, the assessment of tree germination
dates provides – with all limitations involved due to the
rather limited sample size – an independent line of evidence
supporting a causal basis for the correlation of higher tree
age with increasing distance from the source of debris-flow
activity (Bollschweiler et al., 2008b).
The age structure of trees sampled along the Geisstriftbach
torrent also indicates that trees are initially growing up
in cohorts, but that only single trees manage to outgrow competitors.
This growth pattern is one of the reasons for the
clumped distribution of even-aged trees along the channel.
Typical succession in the sub-alpine area with predominance
of conifers usually begins with fast-establishing L. decidua
on raw soil (Frehner et al., 2005). Protected from gliding and
sliding snow, P. abies then begins to regenerate and gradually
replaces L. decidua if growth conditions are stable (Frehner
et al., 2005). However, this change of tree species only partially
occurs along Geisstriftbach torrent as succession is
repeatedly thrown back to the initial stages after devastating
debris-flow events. Under such unsteady growth conditions,
L. decidua should be promoted actively, as it is a tree species
highly adapted to external disturbances (Frehner et al., 2005).
The analysis of the growth disturbances of 26 heavily
affected L. decidua and 2 P. abies allowed for an identification
of 13 debris flows along the Geisstriftbach torrent for
the period AD 1913–2006. But we need to bear in mind that
the debris-flow frequency provided in this study only represents
the minimum number of events that have occurred in
the Geisstriftbach torrent in the period investigated, because
(i) the number of trees selected is rather small, (ii) because
we exclusively sampled trees along the torrent and thus
may have overlooked part of the events and as (iii) individual
events may have been contained within the channel,
i.e. without disturbing any trees in the levees and on the
cone. Similarly, the time elapsed between individual events is
quite regular and does not provide any clear evidence for an
increase or decrease in debris-flow frequency, although such
changes have been put forward as an early indicator of global
change (Jomelli et al., 2004; Stoffel and Beniston, 2006;
Bollschweiler and Stoffel, in press). Due to the restricted
sample depth, implications for forest dynamics are limited.
Lastly, the position of repeatedly disturbed trees provides
evidence regarding preferential breakout locations of debris
flows along the torrent. For example, due to frequent disturbances
reconstructed in the trees sampled at ∼1280 m a.s.l.
along the southern levee, this area can be considered to be
such a preferential breakout location. Even though disturbance
intensity is higher near the cone apex, trees in the
lower part along the Geisstriftbach torrent have also been